Problem: Output of Files with blanks, causes Line break

[raoulvb]

Hi Folks,

i have a strange problem:

try something like that:

create a directory which contains files

file1.txt
file2.txt
file3.txt
i am a file with blanks.txt

then try this

find $SOURCE -type f > /var/tmp/tmparchfiles

for FILE in `cat /var/tmp/tmparchfiles`
do
echo $FILE
done

you will see, that the output will be something like this:

./yourdir/file1.txt
./yourdir/file2.txt
./yourdir/file3.txt
./yourdir/i
am
a
file
with
blanks.txt

not as ihoped:

./yourdir/file1.txt
./yourdir/file2.txt
./yourdir/file3.txt
./yourdir/i am a file with blanks.txt

this is very uncool, because the filename is "broken"

this is already one of my workarounds, the first script looked like this:

for FILE in `find $SOURCE -type f`
do
echo $FILE
done

but this is the same

even with print its the same, do you have an idea how to get the names "without" the line brake?


Thank you in advance

Ra

[burschik]

This behaviour is expected. Use quotation marks (like this: "$VAR") to suppress it.

[raoulvb]

even with quotationmarks, you have the same output

[tetsujin]

Hi Folks,

i have a strange problem:

try something like that:

create a directory which contains files

file1.txt
file2.txt
file3.txt
i am a file with blanks.txt

then try this

find $SOURCE -type f > /var/tmp/tmparchfiles

for FILE in `cat /var/tmp/tmparchfiles`
do
echo $FILE
done

Here's a way to do it:

cat /var/tmp/tmparchfiles | xargs --delim '\n' echo

Of course "xargs" collects arguments from its input, so that it can deliver a large number of them to the command it's supposed to run all at once (so, for instance, "find | xargs rm" could find 10,000 files but only actually invoke "rm" a few times...) This may not be what you want. In this case:

cat /var/tmp/tmparchfiles | xargs --delim '\n' -n 1 echo

(This is a "useless use of cat" of course, but somehow it doesn't bother me. )


Another way to do it is to open the file within the shell, and explicitly read a line at a time from it:

exec 7</var/tmp/tmparchfiles #File descriptor 7 is now open for reading
while read f <&7; do echo $f; done
exec 7<&- #close file descriptor 7


Another option is to change the value of the $IFS variable so Bash will split arguments on newlines rather than on any whitespace character:

IFS="\n"
for f in $(cat /var/tmp/tmparchfiles); do echo $f; done

If you go this route, you'll probably want to change $IFS back to its old value immediately after doing this...


This sort of thing is one of my biggest pet peeves about the Unix shell. There's not a clear-cut way of delimiting multiple, successive items to be processed. Programs commonly separate values with newlines, but the shell separates things with whitespace...

[secondmouse]

I would go for the approach without the temporary file list:

find $SOURCE -type f -exec echo "{}" \;

[raoulvb]

Here's a way to do it:

cat /var/tmp/tmparchfiles | xargs --delim '\n' echo

Of course "xargs" collects arguments from its input, so that it can deliver a large number of them to the command it's supposed to run all at once (so, for instance, "find | xargs rm" could find 10,000 files but only actually invoke "rm" a few times...) This may not be what you want. In this case:

cat /var/tmp/tmparchfiles | xargs --delim '\n' -n 1 echo

(This is a "useless use of cat" of course, but somehow it doesn't bother me. )


Another way to do it is to open the file within the shell, and explicitly read a line at a time from it:

exec 7</var/tmp/tmparchfiles #File descriptor 7 is now open for reading
while read f <&7; do echo $f; done
exec 7<&- #close file descriptor 7


Another option is to change the value of the $IFS variable so Bash will split arguments on newlines rather than on any whitespace character:

IFS="\n"
for f in $(cat /var/tmp/tmparchfiles); do echo $f; done

If you go this route, you'll probably want to change $IFS back to its old value immediately after doing this...


This sort of thing is one of my biggest pet peeves about the Unix shell. There's not a clear-cut way of delimiting multiple, successive items to be processed. Programs commonly separate values with newlines, but the shell separates things with whitespace...

Thank you alot, the underlined one works great for me!

very nice

@secondmouse

im sorry, this doesn´t work for me, because i have several functions that needs the filename, so execute would only work if i outsource this to a second script.

Just for your information, this script is used to perform an archive operation, with mysql documentation in a fully normalized database

thank you alot, for the kind support.

Ra